Large language models (LLMs) trained on massive text corpora may inadvertently memorize sensitive or copyrighted content, motivating the need for more targeted unlearning. Selective LLM unlearning focuses on identifying token-level or span-level unlearning targets within a text, rather than treating entire sequences as unlearning targets. However, many existing selective approaches depend on external supervision to identify unlearning targets, which may misalign unlearning objectives with the model’s internal behavior. In this paper, we propose a selective span-level unlearning method that is grounded entirely in model-intrinsic information. Our method first estimates token-level importance scores by contrasting gradient information induced by forget and retain datasets, identifying tokens that disproportionately contribute to information targeted for unlearning. These token-level importance scores are then used as anchors to identify coherent span-level unlearning targets via a self-consistency–based generation process, allowing the model to determine stable spans based on its own predictions. Experiments on two LLM unlearning benchmarks show that our approach achieves comparable unlearning performance while substantially better preserving retained knowledge.
论文检索
输入标题、作者或关键词，从 12,319 篇学术成果中精准定位
Translation or Recitation? Calibrating Evaluation Scores for Machine Translation of Extremely Low-Resource Languages
PDF ↗The landscape of extremely low-resource machine translation (MT) is characterized by perplexing variability in reported performance, often making results across different language pairs difficult to contextualize. For researchers focused on specific language groups—such as ancient languages—it is nearly impossible to determine if breakthroughs reported in other contexts (e.g., African or American languages) result from superior methodologies or are merely artifacts of benchmark collection. To address this, we introduce the FRED Difficulty Metrics—Fertility Ratio (F), Retrieval Proxy (R) Pre-training Exposure (E) and Corpus Diversity (D) —that serve as dataset-intrinsic metrics to contextualize reported scores. Our findings reveal that a significant portion of result variability is explained by train-test overlap and pre-training exposure rather than model capability. Additionally, we identify that underperforming XLR languages—particularly extinct and non-Latin indigenous languages—suffer from poor tokenization coverage (high token fertility), highlighting structural limitations of transfer learning for languages outside pre-trained models’ representation space. By providing these indices alongside performance scores, we enable more transparent evaluation of cross-lingual transfer and provide a more reliable foundation for the XLR MT community.
Enhancing Linguistic Competence of Language Models through Pre-training with Language Learning Tasks
PDF ↗Language models (LMs) are pre-trained on raw text datasets to generate text sequences token-by-token. While this approach facilitates the learning of world knowledge and reasoning, it does not explicitly optimize for linguistic competence. To bridge this gap, we propose L2T, a pre-training framework integrating Language Learning Tasks alongside standard next-token prediction. Inspired by human language acquisition, L2T transforms raw text into structured input-output pairs to provide explicit linguistic stimulation. Pre-training LMs on a mixture of raw text and L2T data not only improves overall performance on linguistic competence benchmarks but accelerates their acquisition, while maintaining competitive performance on general reasoning tasks.
Retrieval-augmented generation (RAG) over long documents typically involves splitting the text into smaller chunks, which serve as the basic units for retrieval. However, due to dependencies across the original document, contextual information is often essential for accurately interpreting each chunk. To address this, prior work has explored encoding longer context windows to produce embeddings for longer chunks, yet their gains in retrieval and downstream tasks remain limited. This is because (1) longer chunks strain the capacity of embedding models due to the increased amount of information they must encode, and (2) many real-world applications still require returning localized evidence due to constraints on model or human bandwidth. To this end, we propose an alternative approach to this challenge by representing short chunks in a way that is conditioned on a broader context window to enhance retrieval performance – i.e., situating a chunk’s meaning within its context. We further show that existing embedding models are not well-equipped to encode such situated context effectively, and thus introduce a new training paradigm and develop the first situated embedding model. To evaluate our method, we curate a book-plot retrieval dataset specifically designed to assess situated retrieval capabilities. On this benchmark, our 1B-parameter model substantially outperforms state-of-the-art embedding models, including several with up to 7B parameters.
Given a map description through global traversal navigation instructions, an LLM can often infer the implicit spatial layout and answer user queries by providing shortest paths. However, such context-dependent querying becomes incapable as environments grow larger, motivating the need for incremental map construction that builds a complete topological graph from stepwise observations. We propose a framework for LLM-driven construction and map repair, designed to detect, localize, and correct structural inconsistencies in incrementally constructed navigation graphs. Central to our method is the Version Control, which records the full history of graph edits and their source observations, enabling fine-grained rollback, conflict tracing, and repair evaluation. We further introduce an Edge Impact Score to prioritize minimal-cost repairs based on structural reachability, path usage, and conflict propagation. To properly evaluate our approach, we create a refined version of the MANGO benchmark dataset by systematically removing non-topological actions and inherent structural conflicts, providing a cleaner testbed for LLM-driven construction and map repair. Our approach significantly improves map correctness and robustness, especially in scenarios with entangled or chained inconsistencies. Our results highlight the importance of introspective, history-aware repair mechanisms for maintaining coherent spatial memory in LLM agents.
Macaron: Controlled, Human-Written Benchmark for Multilingual and Multicultural Reasoning via Template-Filling
PDF ↗Multilingual benchmarks rarely test reasoning over culturally grounded premises: translated datasets keep English-centric scenarios, while culture-first datasets often lack control over the reasoning required. We propose Macaron, a template-first benchmark that factorizes reasoning type and cultural aspect across question languages. Using 100 language-agnostic templates that cover 7 reasoning types, 22 cultural aspects, native annotators create scenario-aligned English and local-language multiple-choice questions, and systematically derived True/False questions. Macaron contains 11,862 instances spanning 20 countries/cultural contexts, 10 scripts, and 20 languages and dialects (including low-resource ones like Amharic, Yoruba, Zulu, Kyrgyz, and some Arabic dialects). In zero-shot evaluation of 21 multilingual LLMs, reasoning-mode models achieve the strongest performance (80.8% overall) and near-parity between English and local languages (∆MC = −1.3%), while open-weight models degrade substantially in local languages (∆MC = −6.8%) and often approach chance on T/F tasks. Culture-grounded mathematical and counting templates are consistently the hardest. The data can be accessed here https://huggingface.co/datasets/AlaaAhmed2444/Macaron.
Look Within or Beyond? A Theoretical Comparison Between Parameter-Efficient and Full Fine-Tuning
PDF ↗Parameter-Efficient Fine-Tuning (PEFT) has become a popular alternative to Full-Parameter Fine-Tuning (FFT), achieving similar performance on many benchmarks with far lower computational and memory costs. Yet, its effectiveness on complex tasks such as reasoning and instruction-following remains unclear. In this work, we provide a theoretical and empirical comparison of PEFT and FFT in terms of representational capacity and robustness. We show that PEFT’s solution space is a strict subset of FFT’s and derive upper bounds revealing how its restricted parameterization limits expressiveness and increases vulnerability to perturbations. Experiments on 20 datasets and 11 adversarial test sets support these findings, indicating that while PEFT performs well on standard tasks, its weaknesses on complex and adversarial settings call for new directions beyond current PEFT paradigms.The source code is in the anonymous GitHub repository[https://anonymous.4open.science/r/PEFTEval-E2AC ].
Despite significant advances in LLM-driven GUI agents, the field remains constrained by the challenge of reconciling high-fidelity realism with verifiable evaluation accuracy. To address this, we introduce NaturalGAIA, a verifiable evaluation dataset grounded in real-world human GUI interaction intents. By decoupling logical causal pathways from linguistic narratives, it rigorously simulates natural human intent, characterized by cognitive non-linearity and contextual dependencies. Furthermore, we propose LightManus-Jarvis, a hierarchical collaborative framework where LightManus manages dynamic topological planning and context evolution, while Jarvis ensures execution precision via hybrid visual-structural perception. Experiments demonstrate that our approach achieves a Weighted Pathway Success Rate of 45.6%, significantly outperforming the state-of-the-art baseline (21.1%), while reducing token consumption by 75% and execution time by 76%. These results validate the efficacy of the macro-planning and micro-execution paradigm in handling complex naturalized tasks. Our code is publicly available at: https://anonymous.4open.science/r/NatureGAIA-721F/.
Large Language Models (LLMs) can learn both useful knowledge and harmful stereotypes, making bias evaluation essential.Existing frameworks fall into two types: those considering reasoning steps (Thinking Process-Aware Evaluation, TPAE) and those focusing only on final outputs (Straight-to-the-Answer Evaluation, SAE).Prior TPAE studies showed effectiveness in assessing gender bias but relied on template-based, word-counting prompts, limiting generalization to other bias types, languages, and reasoning-based methods.In this study, we introduce MBTP, a multilingual social bias benchmark that incorporates human-generated pro- and anti-stereotype reasoning as part of the thinking process, and propose a few-shot meta-evaluation method that enables scalable bias assessment without model fine-tuning.From experiments evaluating 13 social bias categories across 8 languages, we find that human-generated thinking consistently yields higher-quality evaluations than LLM-generated or template-based approaches.Furthermore, TPAE demonstrates superior performance over SAE, highlighting the importance of considering reasoning processes in bias evaluation.We will release the MBTP dataset upon paper acceptance.
Retrieving music using natural language descriptions has improved with contrastive audio–text models such as CLAP, but current systems remain limited to coarse semantic queries. When descriptions specify fine-grained musical attributes such as tempo, key, chord progression, or rhythmic structure, existing models often fail to retrieve the correct audio. We show that this limitation stems from the contrastive learning objective itself: despite being trained on long captions, CLAP-based models effectively utilize only the first few tokens, discarding much of the information encoded in detailed prompts. Then, we propose FIGMA (Fine-Grained Music Retrieval), a multi-view contrastive architecture that addresses this limitation by jointly optimizing global audio–text alignment and frame-level, token-wise alignment. This design enables FIGMA to capture both high-level semantic context and fine-grained musical attributes within a unified representation space. Moreover, we formalize the task of Fine-Grained Music Retrieval and construct Fine-Grained Music Caption dataset (FGMCaps), a large-scale dataset of 380K music–caption pairs for training along with a 10K test set, both annotated with tempo, key, chord progression, beat count, as well as genre and mood. Extensive experiments demonstrate that FIGMA consistently outperforms existing CLAP-based music retrieval models across multiple music retrieval benchmarks, including out-of-domain evaluations, with relative improvements of up to 73.3%.
GitChameleon 2.0: Evaluating AI Code Generation Against Python Library Version Incompatibilities
PDF ↗The rapid evolution of software libraries poses a considerable hurdle for code generation, necessitating continuous adaptation to frequent version updates while preserving backward compatibility. While existing code evolution benchmarks provide valuable insights, they typically lack execution-based evaluation for generating code compliant with specific library versions. To address this, we introduce GitChameleon 2.0, a novel, meticulously curated dataset comprising 328 Python code completion problems, each conditioned on specific library versions and accompanied by executable unit tests. GitChameleon 2.0 rigorously evaluates the capacity of contemporary large language models (LLMs), LLM-powered agents, code assistants, and RAG systems to perform version-conditioned code generation that demonstrates functional accuracy through execution. Our extensive evaluations indicate that state-of-the-art systems encounter significant challenges with this task; enterprise models achieving baseline success rates in the 48-51% range, underscoring the intricacy of the problem. By offering an execution-based benchmark emphasizing the dynamic nature of code libraries, GitChameleon 2.0 enables a clearer understanding of this challenge and helps guide the development of more adaptable and dependable AI code generation methods.
Dynamic reasoning trees can help large language models solve complex tasks by explicitly structuring intermediate decisions.However, existing approaches often rely on manually specified subproblems or predefined decomposition patterns, which limits the effectiveness of reasoning and generalization.To solve this problem, we propose SyRA, a hierarchical reasoning framework based on MFR theory that supports the construction of adaptive reasoning trees and reliable error correction within a single LLM. Specifically, SyRA focuses on reasoning-tree construction, dynamically controlling branching and expansion using MFR principles to enable informative, non-redundant subproblem decomposition. In addition, it introduces a residual backtracking mechanism for adaptive cross-layer error correction, allowing the model to revise earlier reasoning decisions based on downstream feedback.Across eight reasoning benchmarks, SyRA significantly reduces logical errors and improves reasoning accuracy, while achieving a better balance between accuracy and reasoning time than the Chain-of-Thought, Decompose–Analyze–Rethink and Tree-of-Thought. Our code and dataset are available at https://github.com/Tim798-art/SyRA/tree/main/SyRA.
Visual segmentation, the task of segmenting an image into semantically meaningful regions, is a cornerstone in machine learning and has widespread applications in industry. Nevertheless, visual segmentation with instruction has been a challenging task for many years. This largely stems from the cross-modal discrepancy between language and image domains, resulting in difficulty in relating the instruction semantics and the pixel-level predictions. In recent years, the remarkable reasoning capabilities of Large Language Models (LLMs) and Large Multimodal Models (LMMs) have spurred a new wave of research aiming to bridge the disparity between natural language instructions and pixel-level understanding. This survey offers the first comprehensive overview of the rapidly evolving field of LLM-driven visual segmentation. We categorize existing approaches based on their core objectives and methodologies, including reasoning-based segmentation, open-vocabulary segmentation, grounding techniques connecting language to pixels, and extensions to video domains. We review recent seminal works in LLM-based visual segmentation, analyzing their architectural innovations, training strategies, and benchmark performance. Furthermore, we discuss the common datasets, evaluation metrics, and identify key challenges and promising future directions at the intersection of language and visual segmentation. We hope this survey serves as a valuable resource for researchers and practitioners seeking to understand the current landscape and future directions of leveraging LLMs for sophisticated visual segmentation tasks and applications. The resource summary is available at https://github.com/wyzjack/Awesome-LLM-Visual-Segmentation.
Emotional Support Conversation requires not only affective expression but also grounded instrumental support to provide trustworthy guidance. However, existing ESC systems and benchmarks largely focus on affective support in text-only settings, overlooking how external tools can enable factual grounding and reduce hallucination in multi-turn emotional support. We introduce **TEA-Bench**, the first interactive benchmark for evaluating tool-augmented agents in ESC, featuring realistic emotional scenarios, an MCP-style tool environment, and process-level metrics that jointly assess the quality and factual grounding of emotional support. Experiments on nine LLMs show that tool augmentation generally improves emotional support quality and reduces hallucination, but the gains are strongly capacity-dependent: stronger models use tools more selectively and effectively, while weaker models benefit only marginally. We further release **TEA-Dialog**, a dataset of tool-enhanced ESC dialogues, and find that supervised fine-tuning improves in-distribution support but generalizes poorly. Our results underscore the importance of tool use in building reliable emotional support agents.
FactCorrector: A Graph-Inspired Approach to Long-Form Factuality Correction of Large Language Models
PDF ↗Large language models (LLMs) are widely used in knowledge-intensive applications but often generate factually incorrect responses. A promising approach to rectify these flaws is correcting LLMs using feedback. Therefore, in this paper, we introduce FactCorrector, a new post-hoc correction method that adapts across domains without retraining and leverages structured feedback about the factuality of the original response to generate a correction. To support rigorous evaluations of factuality correction methods, we also develop the VELI5 benchmark, a novel dataset containing systematically injected factual errors and ground-truth corrections. Experiments on VELI5 and several popular long-form factuality datasets show that the FactCorrector approach significantly improves factual precision while preserving relevance, outperforming strong baselines. We release our code at https://ibm.biz/factcorrector.
From Recognition to Reasoning: Benchmarking and Enhancing MLLMs on Real-World Receipt Document Understanding
PDF ↗Extracting structured information from visual documents (Visual Information Extraction, VIE) is a cornerstone of business automation. While recent Multimodal Large Language Models (MLLMs) have shown promising capabilities, existing benchmarks suffer from critical limitations in scale and realism, lack semantic granularity, and fail to cover diverse document types. To bridge this gap, we introduce ReceiptBench, a large-scale, human-annotated benchmark consisting of 10k diverse receipts, organizing information extraction into four hierarchical sub-tasks: (1) Basic Perception for raw text spotting, (2) Format Normalization for strictly following standardization instructions, (3) Semantic Reasoning for inferring implicit attributes from context, and (4) Structure Parsing for handling nested line items. Furthermore, we propose a two-stage training framework incorporating Metric-Aware Group Relative Policy Optimization (GRPO), which translates rigorous evaluation constraints into reinforcement learning signals to enhance structural consistency. Extensive experiments demonstrate that our method yields state-of-the-art performance, surpassing leading proprietary models on complex reasoning tasks. We release our datasets and code at https://github.com/wwwT0ri/ReceiptBench.
Language models are increasingly deployed in interactive settings where users reason about facts over time rather than in isolation.In such scenarios, correct behavior requires models to maintain and update implicit temporal assumptions established earlier in a conversation. We study this challenge through the lens of temporal scope stability: the ability to preserve, override, or transfer time-scoped factual context across dialogue turns. We introduce ChronoScope, a large-scale diagnostic benchmark designed to isolate temporal scope behavior in controlled multi-turn interactions, comprising over one million deterministically generated question chains grounded in Wikidata. ChronoScope evaluates whether models can correctly retain inferred temporal scope when follow-up questions omit explicit time references, spanning implicit carryover, explicit scope switching, cross-entity transfer, and longer temporal trajectories.Through extensive evaluation of state-of-the-art language models, we find that temporal scope stability is frequently violated in controlled multi-turn settings, with models often drifting toward present-day assumptions despite correct underlying knowledge.These failures intensify with interaction length and persist even under oracle context conditions, revealing a gap between single-turn factual accuracy and coherent temporal reasoning under sequential interaction.We make our dataset and evaluation suite publicly available at https://github.com/yashkumaratri/ChronoScope.
Dense retrieval has become a core technique in applications like web search and retrieval-augmented generation. Despite their empirical success, it remains unclear whether these models truly understand semantics. To address this gap, this paper conducts a systematic investigation by introducing SURE, a benchmark for Semantic Understanding in dense REtrieval built upon the MSMARCO, NQ, and FiQA datasets. SURE characterizes semantic understanding in dense retrieval along three dimensions: semantic precision, semantic abstraction, and semantic equivalence. We evaluate ten representative models ranging from 110M to 8B parameters, including both general-purpose and domain-specific models. Results show that current dense retrievers struggle to distinguish fine-grained semantic differences across texts with varying information density, and to recognize semantic consistency under lexical paraphrasing. Moreover, larger models do not necessarily exhibit stronger semantic understanding, and diverse training data generally enhances semantic understanding on challenging retrieval tasks.
Thinking-Based Non-Thinking: Solving the Reward Hacking Problem in Training Hybrid Reasoning Models via Reinforcement Learning
PDF ↗Large reasoning models (LRMs) have attracted much attention due to their exceptional performance. However, their performance mainly stems from thinking, a long Chain of Thought (CoT), which significantly increase computational overhead. To address this overthinking problem, existing work focuses on using reinforcement learning (RL) to train hybrid reasoning models that automatically decide whether to engage in thinking or not based on the complexity of the query. Unfortunately, using RL will suffer the the reward hacking problem, e.g., the model engages in thinking but is judged as not doing so, resulting in incorrect rewards.To mitigate this problem, existing works either employ supervised fine-tuning (SFT), which incurs high computational costs, or enforce uniform token limits on non-thinking responses, which yields limited mitigation of the problem.In this paper, we propose Thinking-Based Non-Thinking (TNT). It does not employ SFT, and sets different maximum token usage for responses not using thinking across various queries by leveraging information from the solution component of the responses using thinking. Experiments on five mathematical benchmarks demonstrate that TNT reduces token usage by around 50\\%$ compared to DeepSeek-R1-Distill-Qwen-1.5B/7B and DeepScaleR-1.5B, while significantly improving accuracy. In fact, TNT achieves the optimal trade-off between accuracy and efficiency among all tested methods. Additionally, the probability of reward hacking problem in TNT’s responses, which are classified as not using thinking, remains below $10\\%$ across all tested datasets.
We introduce SkMTEB, the first comprehensive MTEB-style text embedding benchmark for Slovak, a low-resource West Slavic language, comprising 31 datasets across 7 task types—nearly 4\times the depth of existing multilingual benchmark coverage for Slovak. Our evaluation of 31 embedding models reveals that large instruction-tuned multilingual models achieve the strongest performance, while existing Slovak-specific models trained for NLU tasks transfer poorly to embedding tasks. To address the need for efficient, locally-deployable Slovak embeddings, we develop e5-sk-small (45M parameters) and e5-sk-large (365M) by applying vocabulary trimming and fine-tuning to Multilingual E5 models. Despite size reductions of up to 62%, our open-source models achieve competitive performance with proprietary APIs while remaining locally deployable for semantic search and retrieval-augmented generation (RAG). We release the benchmark, models, datasets, and code openly, hoping our approach offers a replicable path for other under-resourced languages.